Issue 10 / Week of 4 to 17 August 2026

There is a table on the BfArM's website listing every compassionate use programme running in Germany. Twenty-eight entries. Lucerastat for Fabry disease, Idorsia. Tolebrutinib for progressive multiple sclerosis, Sanofi. Pelacarsen for atherosclerotic cardiovascular disease, Novartis. Trofinetide for Rett syndrome, Acadia. Fosmanogepix for invasive mould infections, Basilea. Twenty-seven of the twenty-eight name a pharmaceutical company or its European representative as the responsible party.

The twenty-eighth reads:

Wirkstoff: Psilocybin. Psilocybin bei therapieresistenter Depression. 18.06.2026 to 17.06.2027. Verantwortliche Person: Zentralinstitut für Seelische Gesundheit, Abteilung für Molekulares Neuroimaging, Prof. Dr. med. Gerhard Gründer, J5, 68159 Mannheim.

No company. A university psychiatric institute and a named clinician, listed in the same column where every other row carries a corporate address and a regulatory affairs contact.

We have spent three issues describing a market that manufacturers are not building. Lilly bought AtaiBeckley for $3.8 billion and wrote no European milestone into the contingent value rights. Europe already pays for clinic-administered psychiatry and does not pay for psychedelics, because no manufacturer has taken one through a benefit assessment. The most advanced psilocybin programme in the world has reached the edge of American approval with no European application on file. Three issues, one finding: the commercial layer is absent.

This register entry shows what is left when it is. Not nothing. A psychiatrist filed the application, and the institute says treatment has been possible since 11 July 2025 for patients with treatment-resistant depression, in exceptional cases, at Mannheim and at the OVID day clinic in Berlin. ZI Mannheim describes it as the first compassionate use programme for psilocybin in the European Union. But look at what that costs in reach. A compassionate use programme is an exception granted for individual patients in justified cases. It is not an authorisation, it does not create a reimbursement pathway, and it does not scale, because the thing making it work is one department's willingness to carry the administrative weight that a company would otherwise carry.

The distinction matters more than it sounds. Every other row on that table represents a firm that intends, eventually, to sell the drug in Germany. Compassionate use is the bridge they build to the patients who cannot wait for the authorisation they are already pursuing. There is no such intention behind the psilocybin row, because there is no firm. The bridge leads to a road nobody is building.

Two limits on what we are telling you, both worth stating plainly.

The register publishes one current term, not a history. It shows 18 June 2026 to 17 June 2027 and nothing before that. Prof. Gründer told us by email that the programme has been extended by a further year, and he has agreed to be named. We looked for a public document saying so, across the BfArM, the institute, the Berlin clinic, the supplier and the German press, and there isn't one. Nor have we found a document giving the date the BfArM granted the programme: the 11 July 2025 above is the institute's own statement of when treatment became possible, which is not the same thing. So the renewal is his account, the term is the register's, and we are not going to weld them together with an inference we cannot support.

The register also does not list treatment sites. It names the party who notified the BfArM, which is why Mannheim appears in that column and Berlin does not. The two designated locations are on the public record elsewhere: both the institute's July 2025 announcement and its standing patient information page name the Central Institute of Mental Health and the day clinic of the OVID Clinic Berlin. What is not on the public record is whether both are dosing today. Prof. Gründer tells us they are, and that is his account rather than the register's.

Read the register in one sitting and the psilocybin row stops looking like an anomaly and starts looking like a diagnosis.

The other twenty-seven entries all describe the same shape. A company has a drug in late development or under review, patients exist who cannot wait, and the company builds an access programme that runs alongside the regulatory work it is already doing. Several are explicit about it. The sepofarsen entry is for patients who completed clinical trials in Germany and would benefit from continuing. The rilzabrutinib entry covers people who finished study EFC17093. These are bridges from a trial to a market, built by the party who wants the market.

German law settles one part of the money question and leaves the larger part open. The Medicines Act requires that a drug supplied through a compassionate use programme be handed over free of charge: kostenlos für eine Anwendung bei Patienten zur Verfügung gestellt werden. So the substance costs the patient nothing, by statute.

The substance was never the expensive part. Psilocybin treatment is not a prescription a patient takes home. It is a dosing session with hours of monitoring by trained staff, and that is what has to be funded. This is the Issue 8 argument arriving at a specific address: we wrote then that Europe already pays for clinic administered psychiatry and has simply never been asked to pay for psychedelics.

What we can tell you about how those hours are funded is thinner than we would like, and we are not going to dress it up. There is no published tariff schedule for this programme. The fullest public account we have found is a June 2026 interview in which clinicians from both sites describe the arrangement: Lea Mertens at Mannheim says inpatient treatment is covered under the standard daily copay while approval for outpatient delivery is still pending, and Andrea Jungaberle at OVID Berlin describes securing a fast track agreement with Germany's largest private insurer. That is two clinicians describing their own funding workarounds to a trade publication. It is not a reimbursement decision, and nobody should read it as one.

Which is the point. When a manufacturer runs a compassionate use programme, the funding question has an owner and a paper trail. Here the answer has to be assembled from an interview, because the people carrying the programme are carrying the administration too. The institute said at the outset that it expects demand to significantly exceed capacity, and Gründer's comparison then was Switzerland, where demand far exceeds available supply. A year on, the same June interview puts seven hundred people on the Mannheim list of registered interest. Nothing in this arrangement grows into a national offer, because nothing in it was designed to. The route from here to routine care runs through an authorisation, and an authorisation runs through a company.

Voyage read out on 12 August, and one of our calls is now half resolved. Definium Therapeutics reported positive topline results from Voyage, the first Phase 3 of DT120 orally disintegrating tablet at 100µg in generalised anxiety disorder. The trial met its primary endpoint. HAM-A fell 11.6 points against 6.2 on placebo at Week 12, a placebo adjusted separation of 5.4 points, p below 0.0001, with a standardised effect size of 0.81. Every key secondary was met. Response was 43 percent against 16, remission 14 percent against 4. Mean baseline HAM-A was 28.4 against 27.4, with 107 patients in each arm.

The session burden is in the release too, and it is the number European payers will care about. Participants were assessed hourly from five hours after dosing, and the average time to meeting the end of session criteria was 6.4 hours, with 92 percent clearing by hour eight. That is the clinic day Issue 8 was about.

The release does not mention Europe, the EMA, or any filing outside the United States. We checked the full text rather than skimming it: the trial population is described as adults in the United States, and the only strings resembling European references are inside the words neuroplasticity and Neuropsychopharmacology.

Panorama, the sister trial carrying the low dose arm and the European sites, is guided to September.

AtaiBeckley now says in a filing that Phase 3 has begun, the registry still shows no Phase 3, and the two filings disagree on the size of one trial. The company's quarterly report, filed 11 August for the period to 30 June, states that "BPL-003 Phase 3 activities in treatment-resistant depression (TRD) have been initiated." As of 17 August, ClinicalTrials.gov carries no Phase 3 record for BPL-003 under any sponsor name we can find, and the phrase first patient appears nowhere in the filing. That is the same gap between initiation language and dosing evidence we described in Issues 7 and 9, now with a filing behind it.

The design, from the company's March programme update: ReConnection-1 is a single dose across three arms, 8 mg, 4 mg and placebo, randomised 2:1:2. ReConnection-2 is a two-dose induction, 8 mg against placebo on Day 1 and Day 15, randomised 1:1. Both carry a 52-week open label extension and both take change from baseline in MADRS at Week 4 as the primary endpoint.

The size is where the two documents part company. In March the company put ReConnection-2 at approximately 300 participants. In the August filing it is approximately 230. ReConnection-1 is approximately 350 in both. Neither document explains the difference, and we have not looked beyond the two filings and the company's news release page, so we are reporting the discrepancy rather than characterising it. A change of roughly seventy patients to a pivotal is not by itself bad news, since it can follow a revised power calculation. It is worth knowing about a trial the company describes as already under way.

The Lilly deal has a date. AtaiBeckley's definitive merger proxy, filed 10 August, sets the shareholder vote for 8 September. The deal has not closed. German antitrust clearance came on 30 July and Lilly filed with the Australian competition regulator on 5 August. We note the German clearance only to head off a misreading: that is competition law, not medicines regulation, and it says nothing about an EMA pathway.

Zero, for the third consecutive reading. No newly registered interventional trial with a site in the EU, EEA, UK or Switzerland was first posted in this window, and the same was true of our 3 August and 8 August readings.

We built the base rate in Issue 9 precisely so we would not have to make anything of this. Across 135 weeks, 91 are zero, and runs of three or more have happened sixteen times. So this is the single most ordinary thing this series does, and we are reporting it as a number rather than as a finding. The reason it appears at all is that a measure you publish only when it is dramatic is not a measure.

A statement about our own coverage first. We did not run our full register audit across the EMA, the MHRA, Swissmedic, the G-BA or Zorginstituut Nederland for this issue. Those checks were last completed on 3 August and are two weeks old. We did run a targeted search for major regulator events in the 4 to 17 August window and it surfaced nothing, which is reassurance rather than a substitute. We are telling you both halves rather than presenting an unaudited fortnight as a quiet one, which is the distinction Issue 9 was built on and it would be cheap to abandon it now.

Germany. The psilocybin compassionate use programme is listed on the BfArM register with a term running to 17 June 2027, verified directly this week. Read the lead for what the register does and does not establish.

United Kingdom. No dedicated psilocybin appraisal is in development at NICE. We searched the register again this week and found nothing for psilocybin or psychedelic, and nothing for MDMA within the last six months. We are wording this more cautiously than we did in Issue 9, because our search this week did not reproduce cleanly enough for us to restate the earlier result in full. The British half of the Issue 8 call is unfalsified on what we can see.

Everywhere else. No full audit this session. See above.

Scoring the open calls. Seven are live, and none has resolved, though one has moved a long way.

Issue 4 is now half resolved and is the nearest scoreable call on the board. Voyage has posted a placebo-adjusted separation of 5.4 points on HAM-A at Week 12, and Panorama is guided to September. We are pinning the scoring rule here, in print, before the number exists, because pinning it afterwards is how a ledger stops being worth keeping. The governing metric is the raw placebo-adjusted difference in HAM-A at Week 12, not the standardised effect size. If Panorama's separation between 100µg and placebo comes in under 5.4 points, we are right. If it comes in at 5.4 or above, we are wrong. We are stating that in magnitude deliberately, because a signed comparison inverts the meaning, and we are choosing the raw difference over the standardised one because the two trials are not the same shape. Voyage randomised 214 patients across two arms. Panorama's registry record lists 245 as actual enrolment against a protocol that allowed up to 375, spread across three arms rather than two. We do not know the allocation ratio, so we are not going to compute a per arm figure and present it as fact. What matters for the call is the direction: the arm being compared with Voyage's 107 is smaller, and a standardised effect size responds to that in a way the raw point difference does not.

Issue 3, that AtaiBeckley reports a first BPL-003 Phase 3 patient dosed in the third quarter, resolves on 30 September, forty four days out. The company now states in a filing that Phase 3 activities have been initiated, which is not a dosing report, and the registry still shows nothing. We are not scoring it early in either direction.

Issue 5 waits on the same registry posting as Issue 3 and remains unscoreable. Issue 6 is open. Issue 7 is open and unfalsified, with the deal unclosed and the vote set for 8 September. Issue 8 is open, and its British half is no longer provisional. Issue 9 is open and was not falsified by the Compass second quarter release, which we read in full on 5 August.

This week's call. No pharmaceutical company will be named as the responsible party on a BfArM compassionate use entry for psilocybin or MDMA before 31 December 2027. This is our read and not anyone's stated plan. The reasoning is the lead: the German programme exists because a clinician built it, the register shows no manufacturer anywhere near this indication, and a company that wanted German access before authorisation had an obvious and cheap route to it that none has taken. If a company appears in that column for either substance before the end of 2027, we are wrong. Falsifiable against a public register that we now check.

Resolve: 31 December 2027, or earlier if an entry appears.

If something here is wrong, please tell me, and I will correct it.

Thank you so much for reading.

Dhruv Shekhawat
Founder, Psychedelic Brief